132 RESEARCH WORK [Proc. SOC. Anal. Chem. Research Work in Universities and Colleges of Advanced Technology The following are summaries oi the eight papers presented at the Ordinary Meeting of the Society held in Bradford on Wednesday May Sth 1965 and of the discussions that followed their presentation. The meeting has been recorded in the May issue of Proceedivtgs (Pa 71). Rhodizonic Acid as an Analytical Reagent R-ri R. A. CHALMERS ANI) G. Jf. TELLING" (Cheunistvy Depavtvnent Univevsity of Abevdeen Old A bevdeen) MIL TELLING said that rhodizonic acid had never been isolated but its sodium and potassium salts had been identified in 1825. Sodium rhodizonate was used as a spot-test reagent for barium lead and strontium in neutral or acidic solution and for calcium in the presence of ammonia.Its aqueous solution was unstable especially in direct sunlight. Work had shown that this instability was due to the oxidation of rhodizonic acid to triquinoyl and then to croconic and leuconic acids. This oxidation was catalysed by ultraviolet light and by traces of transition metals. The decomposition rate was retarded but not completely checked by the presence of acetone alcohol or EDTA and by the use of de-mineralised water. Maximum stability was attained between pH 5-05 and 6.5. A satisfactory working solution had been prepared by dissolving 0-21 g of sodium rhodizonate in a 20 per cent. solution of acetone in de-mineralised water. This solution had been prepared freshly each day. In 1942 F. Feigl and H. A. Suter (Ind. Evtg. Chem. Anal. Ed. 1942 14 840) reviewed the existing rhodizonate spot-test methods for barium strontium and lead and extended the test to cover several other uni- and bi-valent metals.The test failed for Ni2+ Co2+ Mn2+ Re2+ and Ca2+ in neutral or acidic solution Fe3+ gave a purple colour believed to be due to the phenolic nature of rhodizonic acid. A quantitative study of these colour reactions had been carried out by first performing the spot-test reactions then diluting to M concentration and plotting the absorption spectra of the coloured solutions. 0 /I /"\ 1 I/ \c/ /I 0 O=C C-0Na O-=C C-ONa (1) The results fell into three groups- (a) Ra2+ Pb2+ and Cu2+ gave fairly stable solutions whose absorption spectra differed considerably from that of the reagent. (b) A@ Sr2+ and Zn2+ gave distinct colours of low stability. (c) The remainder gave the spot-test reaction but after dilution the absorption spectra were identical with that of the reagent.By using a large excess of reagent or by reducing the dielectric constant of the medium with acetone the stabilities of the metal rhodizonates of groups (b) and (c) had been greatly increased. In 95 per cent. acetone solution all the metals formed distinct stable rhodizonates whose colours were very similar to those given in the spot-test reactions. The same principles had been used for examining solutions of those metals that failed to give a colour in the spot-test reaction. For all bivalent metals spectral evidence had been found for the formation of weak complexes or ion pairs between the metal and reagent. Similar evidence had been obtained for the existence of weak complexes of A13+ but Cr3+ failed to react.The spectrum of ferric rhodizonate was very similar to that of ferrous rhodizonate suggesting that the reaction of sodium rhodizonate with Fe3+ was not that of a phenol. An attempt to use sodium rhodizonate as a titrant for small amounts of barium had met with only moderate success. The orange - red solution of barium rhodizonate was essentially colloidal in nature and trouble had been encountered in the formation of particles of irregular size even when gelatin or gum had been used as an anticoagulating agent. * Present address Unilever Research Laboratory Colworth House Sharnbrook Bedford. September 1965) RESEARCH WORK 133 The reagent had been used for determining lead in concentrations in the range 0.3 to 30 p.p.m. The lead solution was mixed with 0.1 per cent.gelatin solution buffered to pH 4-05 to 5.35 and added to a 20 to 30-fold excess of reagent. The volume had been made up to a standard value and after being mixed and set aside for 15 to 30 minutes the optical density had been measured at 570 to 606) mp. Interference from other metals was largely overcome by using this method and it had been found that 1 millimole of lead could be determined with an error of less than 5 per cent. in the presence of equimolar amounts of copper zinc strontium silver cadmium and thallium. Mr. Telling added that- (;) iron must be removed before the determination; (ii) Hg+ should be oxidised to Hg2+ ; and (iii) barium interfered and still presented problems. DISCXJSSION MR. B. KTPLING asked if the stability of the rhodizonic acid solution would be improved by preparing it in de-aerated water and by storing it under an inert atmosphere.Mr. Telling replied that de-aerating with nitrogen had given a slight improvement in stability but storage under an inert atmosphere had not been tried. DR. W. CULE-~)AVIS comniented that rhodizonic acid was unable to form many types of derivative that retained the basic structure of the acid and thus the scope of the investigation was limited. He asked whether for example if one of the phenolic groups were methylated the product would still give the typical reactions of rhodizonic acid. Mr. Telling said that the instability of rhodizonic acid in aqueous solution was so great that no attempt to prepare a derivative had been made. He thought that if one phenol group was methylated the compound would no longer react with bivalent metals.Feigl and Suter (idewz) gave the formation of bivalent rhodizonates as- 0 0 6 0 This would appear to preclude the alteration of one or b.sth of the phenolic groups. MR. C. WHALLEY (Clzaivwzan) enquired how the methods in which rhodizonic acid was used compared with alternative methods already in accepted use for determining the elements mentioned and asked if these methods had any practical advantages over existing ones. Mr. Telling considered that the method for lead might prove suitable for the deter- mination of lead in water samples. The existing methods were based either on the precipitation o f lead sulphide with sodium sulphide in alkaline solution or on extraction with dithizone. Iron would have to be removed as in most methods but the method was more accurate than the sulphide method and less time-consuming than the dithizone extraction method.The barium work would have made an excellent “quick finish’ to sulphate deter- minations but the work was at an impasse at present. The silver complex in 40 to 50 per cent. acetone solution might prove to be the basis of a useful colorimetric method for small amounts of silver. MR H. A. FONER asked if the quality of the reagent was constant from batch to batch from the same manufacturer and whether there was any difference between different manufacturers’ products. In reply Mr. Telling said that most of the work had been done on material obtained from one manufacturer and therefore he had had no real opportunity to compare different manufacturers’ products.The material he had used had shown some variation in solubility. It appeared that most of the material was always readily soluble but a small varying amount had a much lower solubility. Some of this less soluble material had been isolated and examined; its absorption spectrum was identical with that of the more soluble material. Earlier workers had reported that solid sodium rhodizonate existed in two forms and it was supposed that these coincided with the two forms found in the authors’ work. 134 RESEARCH WORK [Proc. Soc. Anal. Chem. The Modification of a Thermobalance to Perform Simultaneous Thermogravimetric and Differential Thermal Analysis BY E. L. CHARSLEY AND J. P. REDFERN (Battevsea College of Technology London S. W. 11) MR. CHARSLEY stated that the two chief methods of thermal analysis were thermogravimetric analysis (TGA) and differential thermal analysis (DTA) in which changes in weight and heat content respectively were measured as a system was heated at a uniform rate.The two techniques were complementary TGA providing quantitative information on processes taking place with a change in weight while YTA could in addition detect such phenomena as fusions and crystalline transitions but was not intrinsically quantitative. The techniques were normally performed with separate instruments and there was oiten difficulty in corre- lating the two sets of results especially in reactions involving a large number of stages. It was therefore advantageous to have a simultaneous record of both weight and heat changes from a single sample particularly when a survey study of a number of compounds was being made.It was the intention in the work described to provide DTA as a supplement to TGA without substantially altering the accuracy of the TGA record. The modifications were made to a Stanton HT-D thermobalance which had a sensitivity of 0.0001 g and a maximum operating temperature of 1400" C. The sample and reference materials were contained in dimped platinum - rhodium crucibles. These fitted into a special ceramic holder seated on top of a 4-bore alumina tube fitted into a socket on the rear balance arm. A platinum platinum - rhodium differential thermocouple system was used with a central platinum - rhodium leg to allow direct measurement of the temperature of the sample or reference material. The three thermocouple wires carried in the alumina rise rod were connected by means of gold-plated plugs and sockets to a bridge fitted to the balance arm.From there connection was made to a fixed terminal block on the balance case through 0-001-inch platinum and platinum - rhodium wires. The AT output was amplified by a Leeds and Northrup d.c. amplifier and fed to a 24-point Bristol Dynamaster recorder giving a maximum sensitivity of 50 pV full-scale deflection. Switching was arranged so that 23/24ths of the time was spent on the AT record. The temperature output was fed directly to the other channel of the recorder which was fitted with cold-junction compensation and calibrated from 0" to 1400" C for full-scale deflec- tion. The holder was shielded from electrical effects caused by furnace switching by means of an earthed nickel sheath introduced between the furnace wall and the conventional ceramic sheath modified for use with dynamic atmospheres.The performance of the thermobalance was found to be unimpaired by the modifications. Owing to the difficulty of obtaining perfectly matched thermocouples a drift of base-line occurred between room temperature and 600" C. The drift was reproducible-hence a correc- tion could be readily made. The performance of the instrument was illustrated by reference to some decomposition curves for inorganic chrctmates and dichromates. A heating rate of 6" C per minute was used with sample weights of 0.1 to 0.2 g. Unless otherwise stated the runs were in static air. The first three weight losses on the curve for magnesium chromate pentahydrate corre- sponded to loss of water as follows- MgCr0,.5H20 -+ MgCrO,.2H20 -+ MgCrO,.H,O -+ MgCrO The last weight loss took place in two equal stages.Each weight loss was closely asso- ciated with an endothermic peak on the DTA record. The anhydrous compound was then stable up to 630"C when it underwent an endothermic decomposition according to the general equation- 2MCr0 -+ 2MO.Cr20 + 1.50,. Zinc chromate could be seen to decompose similarly after loss of water of crystallisation a t 200" C. Although for anhydrous cadmium chromate the TGA record indicated a single- stage decomposition in the range 600" to 720" C with stoiclieiometry corresponding to the equation given above the appearance of two maxima on the DTA record suggested a more complex reaction. September 19651 RESEARCH WORK 135 The DTA curves for sodium potassium and caesium chromates each showed an endo- thermic crystalline transition followed by an endothermic peak due to melting.This was illustrated by the curve for sodium chromate for which peak temperatures for the changes were 430" C and 800" C respectively. Sometimes decomposition of the chromate or dichromate was followed by loss of the metal oxide to give chromic oxide as the final product and this was illustrated by decom- position curves obtained for mercuric chromate and dichromate in an air upflow. The 2-stage nature of the oxide loss was clearly shown on the endothermic DTA peak. Owing to the very high weight losses some distortion of the DTA curve occurred because of vertical motion of the holder in the furnace. The curves for chromium trioxide in dynamic air showed that after fusion of the com- pound at 200" C decomposition to chromic oxide took place in three stages the first two being exothermic.The reversible nature of the decomposition of calcium strontium and barium chromates was illustrated by the analysis of a mixture of calcium chromate and chromic oxide formed by the thermal decomposition of calcium dichromate. In a dynamic air upflow the compound was stable at 1000" C. On introducing an upward flow of nitrogen the chromate began to decompose losing oxygen and there was an associated endothermic peak on the DTA record. On switching off the nitrogen the reaction was reversed with an exothermic weight gain. The authors had found that their modification had produced a system of good sensitivity although it was subject to the limitations mentioned.In future work they hoped to decrease the drift by careful matching of the thermocouple system and to extend the range of the instrument by fitting a platinum shield in place of the present nickel shield. In most decompositions this effect was negligible. DISCUSSION 1 9 ~ . M. WILLIAMS asked if the authors had found any discrepancies between their results for chromates and dichromatcs and those published by Erdey and his co-workers which had been determined with the derivatograph. Mr. Charsley replied that of the compounds studied by Erdey and co-workers only caesium chromate had been studied in their apparatus. Their results were essentially the same although they had found a 2 per cent. weight loss below 1000° C and their peak temperatures of 770" C and 955" C were lower than those found by Erdey (910" C and 1050" C).MR. C. WHALLEY (Chaivman) asked if zinc tetroxychromates had been examined since he thought i t would be interesting to compare their compositions with those of the normal chromates. The decomposition of the basic chromate should be catalysed by the presence of zinc oxide. Dn. W. CULE-~)AVIES asked if the products arising from the decomposition of chromic acid (CrO,) by heat had been identified and related to the indications given on the differential thermal analysis and thermogravimetric analysis curves. Mr. Charsley replied that no studies had yet been made on the intermediates formed in this reaction. Previously published papers suggested that it was very difficult to make an exact deter- mination of the nature of these intermediates.MK. B. S. COOPER emphasised the value of supplementing results obtained by thermal techniques with information from additional investigations. In particular X-ray diffraction studies could provide tangible and exact evidence of the nature of the phase changes. He asked about the nature of the atmosphere used in the experiments described by the speaker. In order to characterise a decomposition system completely i t was necessary to apply a variety of techniques such as chemical analysis infrared spectroscopy X-ray diffraction and magnetic studies. In their survey they had mainly used the first two techniques and they intended to use the last two for detailed investigation of the reactions of particular interest. Runs in dynamic atmospheres had been made with a special device to allow controlled upflow of air or nitrogen.Mr. Charsley said that only the normal chromate had been studied. Mr. Charsley agreed that these observations were very true. 136 RESEARCH WORK [Yroc. SOC. A~zal. Chew. Studies on Inorganic Systems of Analytical Interest Scandium(II1) 8-Hydroxyquinolinates BY T. J. CARDWELL AND R. J. MAGEE (Department of Chemistry Queen's University Belfast) MR. CARUVVELL said that in 1951 Pokras and Bernays had precipitated a lemon-yellow compound for which the formula ScQ,.HQ (where Q = C,H,NO) had been given by conven- tional means with 8-hydroxyquinoline. In the present work scandium(rI1) 8-hydroxy- yuinolinate was precipitated from homogeneous solution with 8-hydroxyquinoline generated by the hydrolysis of 8-acetoxyquinoline.In one scandium was precipitated as a bright yellow compound (SCQ,)~.HQ at pH 8.8 and in the other a lemon yellow compound ScQ,.HQ was precipitated at pH 6.5. Further investigations had shown that a third compound an orange scandium S-hydroxy- quinolinate could be prepared by a solid-phase reaction between the scandium S-hydroxyquin- olinate ScQ3.HQ and 8-hydroxyquinoline. The composition proposed for the orange com- pound was ScQ,. 1.5 HQ. THERMOGRAVIMETRIC ANALYSIS- Previous studies had been made by Wendlandt in 1956 with the conventionally prepared scandium 8-hydroxyyuinolinate. In the present work it had been found that the pyrolysis curves for the complexes from homogeneous solution were similar to those obtained by Wend- landt except for that of the orange compound prepared by solid-phase reaction.Above 125" C the compound pre- pared in the solid phase was least stable. It apparently lost half a molecule of 8-hydroxy- quinoline between 200" C and 270" C forming ScQ,.HQ. In all instances no constant weight level corresponding to the normal chelate ScQ, had been obtained; the remaining organic material was oxidised between 450" 6 and 550" C and conversion into the oxide Sc,O, was complete above 600" C. INFRARED SPECTRAL STUDIES- The spectra of the three complexes were very similar the largest differences occurring in the spectrum of the complex prepared in the solid state. Minor differences occurred in the 12 to 14-,u region indicating only small differences in the structures of the complexes. Small differences occurred also in the long-wavelength region which would appear to indicate that there was some change in the M-N linkage in the orange complex.Two procedures had been developed. Studies on the three compounds had been made by the following methods. All the complexes were thermally stable up to 125" C. No work had been carried out previously with scandium 8-liydroxyyuinolinates. ULTRAVIOLET - VISIBLE ABSORPTION SPECTRA- as follows- 8-Hydroxyquinoline and the three scandium complexes had shown absorption peaks 8-hydroxyquinoline 250 and 315 my ScQ3.HQ (ScQ312J3Q ScQ,.l.SHQ 250 317 and 375 mp 250 317 and 375 mp 250 317 and 375 mp Attempts had been made to interpret the structure from the absorption curves of these compounds on the basis of the theory of orientated light absorption. DIFFUSE REFLECTANCE SPECTRA- had shown peaks as follows- The diffuse reflectance spectra of 8-hydroxyquinoline and the three scandium complexes 8-hydroxyquinoline 240 and 315 mp ScY3.W (ScQJ2.HQ SCQ,.1 *5HQ 240 325 and 375 mp 240 340 and 390 inp 2/40 325 and 390 mp September 19653 RESEARCH WORK 137 A peak-by-peak comparison of the ultraviolet - visible spectra of the three scandium complexes in chloroform with the diffuse reflectance spectra indicated initially small shifts in the peak values. However an over-all comparison of the spectra had shown clearly that appreciable differences were occurring indicating that on dissolution in chloroform all three scandium complexes produced the same species in solution despite the individual differences in the solid state. Further on the basis of eye-response sensitivity the different colours of the three complexes had been explained.DISCUSSION MR. T. R. GRIFFITHS said that the authors had shown a cliffuse reflectance spectrum that included the ultraviolet region and in this region a peak had been obtained with an optical density of 1 . 1 . I t was his experience t h a t peaks in this rcgion coulct not bc relied on unless they were still present after further dilution with optical densities of below 0.6. Further he preferred lithium fluoride as a diluent he asked if the authors had tried this or other diluents to see if there was any diffcrence in the spectra recorded. Slow surface reactions sometimes occurred and he enquired i f the authors had investi- gated the stability of their spectra with time. He referred the authors to the article on diffuse reflectance spectra in the ultraviolet region by T.R. Griffiths K. A. K. Lott and C. R. Symons (Anal. Chem. 1959 31 1338). In reply Mr. Cardwell said the diffuse reflectance spectra had been made with a Hilger & Watts Uvispek. The maximum optical density obtained in the spectra had been 1.1 and the authors had found little difficulty in obtaining optical densities in this range. On further dilution with silica as diluent the peaks in the ultraviolet region still remained. Lithium fluoride had not been used and the spectra showed no change with time. With reference to the reflectance curves for the autliors’ samples MR. E. COATES noticed that optical densities had been plotted against wavelengths. Optical density he said had no particular significance in such instances and i t was usual to consider the percentage reflection or reflectance which was related to the amount of pigment.He asked if there was some special reason for the use of optical density in the works. Mr. Cardwell said there was not; they had only been concerned with the position of the peak in order to explain the structure of the complexes investigated. MR. G. M. TELLING asked why the authors had considered that mixing scandium 8-hydroxyquinolinate with 8-hydroxyquinoline in the dry state should give a new com- plex. The speaker replied that it had been simply by chance that they had investigated the solid-state reaction of scandium 8-hydroxyquinolinate with 8-liydroxyquinoline. Substoicheiometric Determination of Traces of Gold by Neutron-activation Analysis BY I).A. BEARDSLEY G. K. RRISCOE J. RUZICKA AND M. WILLIAMS (College of A duanced Technology Gosta Green Birmingham 4) MR. BEARDSLEY said that the substoicheiometric principle applied to neutron-activation analysis had three main advantages over conventional methods higher selectivity avoidance of determining chemical yield and greater simplicity. Such a highly selective simple and rapid radiochemical separation procedure for gold had been developed. The method was based on the solvent extraction of gold diethyldithio- carbarnate from a medium of mineral acid by use of a substoicheiometric amount of copper diethyldithiocarbamate as ctielating agent in chloroform as solvent (I Extraction could be carried out from 0.05 N to 8 N sulphuric or hydrochloric acid; considerable amounts of other acids (hydrofhoric or nitric) might also be present.From a number of metals studied only palladium interfered if present in greater amounts. The time necessary for reaching extraction equilibrium which ensured the reproducibility of this type of determination was less than 1 minute. 138 RESEARCH WORK [Proc. SOC. Anal. Chem. On the basis of the 1-step procedure outlined above a method of determining traces of gold by neutron-activation analysis had been developed. The results of analyses of high-purity lead and standard rock W l were in good agreement with the findings of other authors. Semiconductor-grade silicon and a vegetable material had also been analysed ; amounts of gold down to 2 x 10-log had been determined. DISCUSSION MR. E. L. CHARSLEY asked what the accuracy and reproducibility of the authors’ method was and how they compared with those of the classical method.Mr. Beardsley replied that the accuracy of the method had been tested by analysing several standard materials (containing about The results were in good agreement with those of other authors who used conventional methods of neutron-activation analysis. The reproducibility of this method was at least that obtainable by conventional neutron- activation analysis the accuracy of which was of the order of A5 per cent. The main advantage of the proposed method was that it was simpler than any previously described activation-analysis method for gold. MR. C. T. WESTON asked if the neutron irradiation was carried out with a neutron source (such as polonium -beryllium) or if an irradiation in a Harwell reactor was necessary or in other words what neutron flux was required.Supposing the method were to be adopted industrially he wished to know what the irradiation would cost per sample. He also asked what weight of sample was needed for the method and since gold had a relatively high cross-section how successful the method would be for materials of very low cross-section. Mr. Beardsley explained that all the irradiations had been carried out either at Harwell (in a flux of 6 x to 6 x 1012 neutrons cm-2 sec-l) or at the Universities of Liverpool and Manchester Research Reactor at Risley (in a flux of 2 x 1011 neutrons crw2 sec-l). For the preliminary studies where a high flux had not been required the Research Reactor had been used. For the analysis of the different matrices the flux required depended on the expected gold content and the particular matrix.A polonium - beryllium neutron source could not be used as the flux was too low for trace analysis. The cost o€ an irradiation at Harwell was dependent on the flux and duration of irradiation; they had used a flux of 6 x 1011 to 6 x 1012 neutrons om-2 see-3 for 4 to 5 clays the cost for 1 standard can being L11 in BEPO and L25 in DIDO. The weight of sample used was also dcpendent on the expected gold content matrix and neutron flux available. Also possible nuclear interferences had to be considered. For the silicon rock and kale samples a 1-g trial was used and €or lead a 2 to 3-g trial Many elements having less favourable cross-sections than gold had been determined by activation analysis (R.C. Koch “Activation Analysis Handbook”). DR. J . B. HEADRIDGE asked if the method could be applied to the determination of trace amounts of other elements with results as good as those for gold. The lecturer replied that the substoicheiometric method of neutron-activation analysis had been developed for 24 different elements with very good results as had been found with gold (J . Ruzicka and J. Stary A tovvzic E:nevgy Review Vol. 2 No. 4). Mr. 33. A. Mi. PRATT asked if any special precautions were necessary for ensuring pro- tection from radiation while carrying out the experiments. Mr. Heardsley said that the normal precautions necessary when activity at the millicurie level was used were all that were required. The radiation from any of their samples after removal from the standard irradiation can was of the order of 1 to 10 milliroentgens hr-l which was very low in comparison with the maximum permissible dose per week of 0.1 roentgen of p or y radiation.p.p.m. of gold). September 19651 The Synthesis of Arose in BY RESEARCH WORK 139 Some New Complexones and a Discussion of the Problems that the Determination of Their Acid Dissociation Constants H. M. N. H. IRVING L. D. PETTIT AND M. G. MILES (Depavtment of lnovga& Chemistry The Uwive&ly Lee& 2) MR. MILES gave an account of the various problems encountered in the preparation and purification of a series of new aminopolycarboxylic acids of the general type (I) and (11) formed from the respective carbonyl compound by conversion to an amino-acid followed by CH,CO,H C €3 ,C 0 H / / (1) (11) condensation with chloroacetic acid under alkaline conditions.In order to determine their stoicheiometric acid dissociation constants and metal stability constants by potentiometric titration in a medium of 0.10 N potassium chloride at 20” C a suitable electrochemical cell that would give hydrogen ion concentrations was needed. Previous workers had used both hydrogen and glass electrodes as the indicator electrode and silver - silver chloride or calomel electrodes as the reference electrode. In the present work particular attention had beerr given to cells (A) and (B)- -G.E. I H+ L C1- 11 KCI sat. 1 Hg2CI2 (s) - Hg (l)+ . . (A) G.E. I H+ L Cl- I/ 0.1 M KC1 11 KCI sat. 1 Hg,Cl (s) - Hg (I)+ . . . (B) salt bridge These cells had been standardised in terms of hydrogen ion activity (H+) by using either 0-05 M potassium hydrogen phthalate (p (H ) = 4.000) OY 0.05 M potassium hydrogen phthalate in 0.05 M potassium chloride (p (H ) --7 3.960).When used in titrations of EDTA and nitrilo- triacetic acid with standard base the calculated values of their acid dissociation constants (and especially those of pK and pK,) had been found to vary considerably depending on the cell used and on the method of standardisation. These variations had been shown to be due to the difference AEi between the liquid-junction potential of the buffer solution and the saturated calomel electrode and that of the test solution and the saturated calomel electrode respectively. Since (H ) = [HI f ~ where (HI) [HI and f~ denoted the activity the concentration arid the activity coefficient respectively of the species H values of p(H) the pH meter reading could be related to the hydrogen ion concentration by the equation- y(H) = p[H] - IOgfH - AEj = p[H] -+ A .The value of A had been found to be different for each cell and for each procedure of stan- dardisation ; and its determination was necessary so that the hydrogen ion concentration [HI could be obtained for the calculation of stoicheiometric acid dissociation constants and metal stability constants. To do this a solution of a standard acid of concentration and ionic background comparable to that of the test solution had been titrated with standard base and A had been given by the difference between the experimental p(H) readings and the calculated values The speaker stressed that great care was necessary when the pH-meter readings ona cell with a liquid junction were interpreted and full details of the procedure recommended for standardising a cell in terms of hydrogen ion concentration were given.DISCUSSION DIL NI. A. LEONARD asked if Nr. Miles had experienced difficulty due to the in- solubility of fully protonated EIITA and similar material and how he would set about determining pK values for a very insoluble polybasic acid. 140 RESEARCH WORK [Proc. Soc. AYtal. Chew. M at 20" C) was sufficiently soluble to allow a value of pK to be measured by potentiometric measurements with reasonable accuracy. But as the solubility of a complexone decreased below this value the accuracy of the pEi value so obtained would decrease ; consequently eithev measure- ments must be performed in a series of mixed solvents in which the acid was more soluble with subsequent extrapolation back to water ov an alternative must be tried viz.a spectrophotometric method by measurements of partition or solubility as a function of pH. Solution in alkali and subsequent back-titration with acid would often give the more useful pK values. MR. C. WHALLEY (Chaivvuzan) expressed interest in the new complexones that had been synthesised and asked how far the authors had progressed in assessing their analytical value. He enquired what advantages they would have over the conventional corn- plexones. The speaker said that two of the ten ligands synthesised had shown an inversion of the usual order of the magnesium slid strontium complex stabilities viz. KsrL > K M ~ L but the other ligands had shown no unusual behaviour and the order of stabilities of the complexones formed with the group HA metals was K c a ~ > K M ~ L > K s r ~ > K B ~ ~ .Also The chelating power of the ligands was generally less than that of nitrilotriacetic acid. However a detailed investigation of possible analytical values had not yet been made. MR. E. CHRRSLEY commented that in applying the equation d G = dH - TdS to the series of complexones the authors assumed a constant value for the entropy of formation and asked how they evaluated this. Mr. Miles replied that for two ligands of very similar stereochemistry the entropy of formation of their complexes with metal cations could in certain circumstances and to a first approximation be assumed to be constant. XJnder these conditions the difference in free energy and hence in the measured stability constant could be directly related to the difference between the heats of forma- tion of the two complexes.The actual evaluation of the entropy of formation did not then arise. MR. I?. COATES said that in the proposed treatment of the thermodynamic data for the equilibria involving metal ions and the complexones the speaker had suggested a constant entropy of formation for the complexones. Since some of the complexones were phenyl derivatives and others were non-aromatic the entropy changes were not likely to be the same for the two classes. He asked if the speaker intended the entropy assumption to apply to the whole group of complexones. No said Mr. Miles it only applied to the phenyl derivatives for which the only variation in the ligand molecule was the P-substituent.In reply Mr. Miles said that EDTA (solubility Development of a Spectrophotonietric Method for the Determination of Lead in Steel Brass and Bronze BY R. M. DAGNALL T. s. WEST AND P. YOUNG (Chemistry Department Imperial College London S. W.7) MR. YOUNG said that lack of stability in existing spectrophotometric methods based on the reaction between lead and dithizone had led to a systematic search for a more reliable pro- cedure. A preliminary investigation of a wide selection of possible metallochromic reagents had shown 4-(2-pyridylazo)resorcinol to be most suitable giving a red-coloured complex with lead in arnmoniacal solution. After preliminary absorption spectra had been obtained optimum conditions for the maximum colour development had been determined.The nature of the complex had been established by three optical procedures and by pstentiometric titration and found to be a 1 to 1 ligand - lead complex. Many cationic interferences had been found of which only a few could be masked with cyanide. A selective solvent separation procedure had been developed based on extraction of lead iodide from dilute acidic solution into isobutyl methyl ketone. Lead had then been back extractecr into an aqueous cyanide solution buffered at pH 10 and the 4-(2-pyridylazo)- resorcinol complex developed. September 19651 RESEARCH WORK 141 The procedure had then been applied successfully to the determination of lead in iron mild steel alloy steels brass and bronze alloys (K. M. Dagnall T. S. West and P. Young TalaPzta 1965 12 589).Ferrous alloys required a preliminary separation of most of the iron by extraction with amyl acetate ; copper-base alloys were treated by preliminary concentration of the lead on freshly precipitated calcium carbonate. DISCUSSION In answer to DR. J. B. HEADRIDGE who asked what was the lowest concentration of lead that could be determined in alloys by using the authors’ method Mr. Young said it was 0.001 per cent. starting from a I-g sample. DR. M. A. LEONARD said that Mr. Young had stated that his result of a 1 to 1 complex at pH 10 conflicted with other results that showed a 2 to 1 complex. He asked if these other workers had established their ratio a t the same pH. Mr. Young said that Pollard and co-workers had reported that lead formed a 1-95 to 1 ligand - lead(I1) complex at pH 10.However Kristiansen et al. had reported a 1 to 1 complex at this same pH. DR. J. M. SKINNER asked if the pH for maximum extraction of the lead back into aqueous solution coincided with the pH for optimum development of colour of the 4-( 2-pyridy1azo)resorcinol complex or if the extraction pH was different from that of maximum colour development. Mr. Young replied that the back extraction of lead into aqueous solution was a maximum at pH 10 which was also the optimum pH value for colour development. MR. T. R. GRIFFITHS said that the authors had quoted an extinction coefficient of 40,000 for their complex. He wanted to know the accuracy of this figure and the mean standard deviation obtained from their various determinations of it. The authors had found a reproducible extinction coefficient of 40,000 consistent with the accuracy in determining absorbances Mr.Young said. However a more useful guide to the accuracy of the method was the value of &0*0007 per cent. for the mean standard deviation obtained from analysis of a steel sample containing 0.0151 per cent. of lead. The Polarographic Determination of Lead in Steels and High-temperature Alloys BY A. G. HAMZA AND J. B. HEADRIDGE (Department of Chemistry The University Shefield 10) MR. HAMZA said that recent work in the Sheffield laboratory had shown that in M ammonium fluoride base electrolyte adjusted to pH 7 lead gave a reversible reduction wave at E+ == -0.453 volt veysus the saturated calomel electrode whereas 34 other ions including those of titanium niobium and molybdenum gave reduction waves well removed from the lead wave or were not reduced at all with the exception of tin@) thallium(1) and tellurium(rv) with half-wave potentials of -0,412 volt (oxidation wave) -0.434 volt and -0.52 volt respectively.Since thallium was unlikely to be present in high-temperature alloys and since it was a simple matter to arrange for tin and tellurium to be present as tin(rv) and tellurium(vI) which were not reduced polarographically in the fluoride base electrolyte a method had been developed for the direct polarographic determination of lead (0.01 to 2 per cent.) in such alloys provided that the iron content did not exceed that of the lead. Unfortunately iron(m) gave an irreversible wave at -0.77 volt versus the saturated calomel electrode and on a d.c. polarogram this reduction wave interfered with the lead wave when the amount of iron was in excess of the lead.For alloys containing more iron than lead the lead was separated from the iron by absorption on a column of Zeo-Karh 225 cation-exchange resin in the hydrogen form from M hydrofluoric acid eluted from the column with 2 M hydrochloric acid and determined polarographically in M hydrochloric acid solution. By this procedure 0.01 to 2 per cent. of lead in steels had been determined with good accuracy and precision. 142 RESEARCH WORK [Proc. SOC. Anal. Chem. DISCUSSION In reply to MR. H. A. FONER who asked what materials had been used to construct the ion-exchange apparatus Mr. Hamza said the column was constructed from polythene tubing. DR. W. CULE DAVIES observed that glass capillaries used as dropping-mercury electrodes were seriously corroded by fluoride ground solutions thus making frequent standardisation necessary.He asked how the authors had overcome this difficulty. Mr. Hamza agreed that ordinary lead-glass capillaries were attacked by fluoride solutions producing a lead wave but said that the corrosion was negligible with Pyrex-glass capillaries at pH 7. UR. J. M. SKTNNER asked wliat the total volume of the final solution for polaro- graphic examination was. i f a 1-g sample containing 0-01 per cent. of lead was used he thought a concentration step might have been necessary after elution to achieve the required sensitivity on the dx. polarograph. Mr. Hamza replied that the total volume of the final solution was 100ml after an ion-exchange separation. For 0.02 per cent.of lead in a 1-g sample this corresponded to a final lead concentration of about 5 x 10-6 M. This concentration of lead could just be determined on the Sargent model XV polarograph when the micro-range extender was used. MR. C. M. TELLING wished to know more about the separation technique used to isolate lead copper etc. from the bulk of the iron etc. and asked whether the separation would work in the presence of an acid other than hydrofluoric. Mr. Hamza said the separation technique was as follows- A solution of the alloy M in hydrofluoric acid was passed through a column of Zeo-Karb 225 in the hydrogen form. Lead copper cobalt nickel and a trace of chromium(II1) were adsorbed and iron and all other ions were washed through with M hydrofluoric acid. All of the lead and most of the adsorbed materials were then eluted with 2 M hydrochloric acid.This separation scheme had been developed by J. R. Headridge and E. J. Dixon (Analyst 1962 87 37) and E. J . Dixon (Ph.D. Thesis Univer- sity of Sheffield 1963). This particular separation of lead from other elements that would interfere in its polarographic determination was not possible in solutions of acids other than hydrofluoric Rapid Automatic Coulometric Titration BY D. ROBINSON (Department of Chemical Technology Bradford Institute of Technology Hradfovd 7) MR. ROBINSON said that the many inherent advantages of titration with electro-generated reagents e.g. elimination of titrant and related dispensing and storage apparatus con- venience and simplicity of operation had not resulted in a correspondingly wide adoption of the technique in analytical laboratories in the United Kingdom.He suggested that this might be a result of a lack of availability of commercial instrumentation and of standard methods and procedures. A suitable combination of a coulometric-titration unit with a dual indicator-electrode amperometric end-point system produced a rapid automatic titrator the “dead-stop” metal electrodes being simpler and more convenient than the reference electrodes of potentiometric titrators. However because of the empirical nature of the end-point system and the absence of any anticipatory or “false-end-point” mechanisms accuracy and precision were often partly sacrificed for speed and simplicity of operation. The E.E.L. chloride meter was one of a number of commercial instruments based on the conlometric principle.It was push-button operated and had a digital read-out. It operated by generating silver ions from a silver anode by means of a constant current in order to precipitate chloride from the sample ; at the equivalence point an excess of silver ion appeared and was detected by an indicating circuit of two silver electrodes and a microammeter relay. The relay stopped a digital counter which directly indicated the chloride concentration in the sample. The generator cathode was also silver and was not isolated. September 19651 RESEARCH WORK 143 The application of the instrument to the determination of chloride and various other anions that precipitated silver ion was examined. Its evaluation for chloride gave an error of 0.5 per cent. depending on the sample size.Although most silver salts were insoluble few were sufficiently insoluble in the strong acid necessary for this procedure. Bromide and iodide gave slightly less accurate results than chloride probably owing to slight reduction of the silver halide a t the cathode which could probably be prevented by its isolation. Thiocyanate gave results similar to those given by chloride and by back-titration of the excess of thiocyanate copper nickel and cobalt could be determined when they were precipitated as their pyridine - thiocyanate complexes. Cyanide gave poor results possibly owing to the volatility of hydrogen cyanide and the moderate solubility of silver cyanide in the acid media. Replacing the silver generating anode and indicator electrodes by platinum and isolating the cathode by using a sintered filter-stick made it possible to use the meter to generate iodine.This permitted the determination of thiosulphite arsenic(II1) and antimony(II1). Tin(I1) gave poor results doubtless owing to aerial oxidation. Unfortunately fast direct reactions with iodine were rare; most analytical reactions of iodine were carried out with an excess which was then back-titrated and this limited the number of applications of automatic coulometric methods with iodine. Studies with the generation of bromine were to be made as this promised a much wider scope and included many organic applications. A useful possibility was the determination of metals with EDTA by using mercury indi- cator electrodes. The lecturer hoped that he had shown that the potential use of a rapid coulometric titrator was very wide and that it could be used much more than it was at present.DISCUSSION DR. M. WILLIAMS said that the speaker had mentioned in connection with the determination of metals as their pyridine - thiocyanates that these precipitates must be isolated before the excess of thiocyanate was back-titrated. We asked how the pre- cipitates were isolated and if it was by filtration. Mr. Robinson replied that in practice the precipitates settled very easily and the excess of thiocyanate was determined by sampling an aliquot from the supernatant solution. MK. A. HEDLEY commented that the advantages of the apparatus for performing large numbers of the same titration were obvious and asked how easy it was to change from one type of titration to another. The lecturer had said that in an automatic method the titration was stopped when the current reached a pre-set value.This value was at a given time after the end-point and this time lag was considered to represent the blank value. Mr. Hedley commented that after the end-point the current did not necessarily rise at the same rate and would therefore reach the pre-set value at various times after the end-point. The built-in blank value therefore would not necessarily be correct. Mr. Robinson replied to the first part of the question by saying that the generating cathode and two indicating electrodes were in the form of a common easily replaceable plug-in unit and the generating anode was simply held by a screw clamp so that electrode exchange was simple. Answering the comment on blank value he said he had no doubt that variation in blank value between sample and blank was one of the factors limiting the accuracy of the method.MR. K. G. POWELL said that Mr. Robinson had suggested that long delay times for the detection of end-points were experienced when the “dead-stop” (biamperometric) method was used. He enquired if the speaker had investigated the use of differential electrolytic potentiometry for end-point detection. E. Bishop (e.g. AmZyst 1958,83 212) had standardised the conditions for several titrimetric systems by using this potentio- metric method which gave excellent end-point detection for manual titrations. Mr. liobinson admitted that the delay or “over-titration” period was several seconds but as it was reproducible it was no great disadvantage. Potentiometric end-point methods had not been investigated as the “dead-stop” method was considered to be simpler especially in an automatic instrument. In general other factors such as stirring lag were probably more limiting than the end-point technique which was considered satisfactory.